Product market fit stages move from finding a real customer problem, to validating a solution, to making demand repeatable. Fit builds in levels rather than arriving as a single yes-or-no moment, so each stage needs its own evidence.
This guide sits within our wider look at how to grow a startup. Use it to decide which stage you are in before you spend on growth.
Why fit builds in levels
Product-market fit is not something you have or lack. First Round says it progresses in levels and is not a binary state in which you have it and are golden. It describes four levels: Nascent, Developing, Strong and Extreme. It also frames fit along three dimensions: Satisfaction, Demand and Efficiency.
The top level is defined by demand. First Round calls extreme product-market fit a state of widespread demand for a product that satisfies a critical need. It names repeatability as the goal, meaning results that hold across demand, product, satisfaction and unit economics.
In practice, you should ask which dimension is weakest, not whether you have fit. A product people love but few discover has a demand gap. A product with strong demand but heavy acquisition cost has an efficiency gap. Each gap points to different work.
The roadmap: three stages and how other models compare
The simplest useful roadmap has three stages: discover the problem, validate the solution, then scale what repeats. It lines up with Productboard’s summary of Lean Startup author Ash Maurya, who breaks the startup lifecycle into three stages: Problem/Solution Fit, Product/Market Fit and Scale.
Other writers cut the journey differently. Zendesk reduces it to two product-market fit stages, before you find fit and after you achieve it. A LinkedIn article by Yann Goarin uses five: Discovery, Validation, Repeatability, Efficiency and Expansion. The labels differ, but all of them separate learning what customers need from proving that the result can be repeated. Pick the granularity that matches your decisions.
| Stage | Question to answer | Signpost to look for | Planning use |
|---|---|---|---|
| Problem discovery | Do we have a problem worth solving? | 50+ customer discovery interviews | Stay here until the pain is clear and the same customer type keeps describing it. |
| Solution validation | Have I built something people want? | 3-5 similar design partners actively using your product | Treat active use by similar customers as the gate to Stage 3. |
| 3. Repeatable growth | Does the pattern repeat across customers? | 5-25 customers with consistent success patterns; Net Revenue Retention approaching or exceeding 100%; CAC payback under 12 months | Scale spend only when retention and payback both hold. |
The questions come from Productboard and the signposts from Goarin’s five-stage model, which are suggested markers rather than universal thresholds.
Stage 1: problem discovery
Customer discovery comes first: finding a problem painful enough that a specific group will pay to solve it. The Goarin model suggests conducting 50+ deep customer discovery interviews in this stage. Treat that as a guide to depth, not a pass mark. The point is to hear the same problem described repeatedly by one customer type.
Product manager Dan Olsen’s six-step method, summarised by ProductPlan, makes the same move. After choosing a target customer, its second step is to identify that customer’s underserved needs. Only later does it specify a minimum viable product feature set. Building before that step is the most common way to skip the stage.
Move to Stage 2 once you can name the customer, the underserved need and the evidence behind both. If the interviews keep producing different problems, narrow the customer segment before building anything.
Stage 2: solution validation
Solution validation tests whether a minimal product earns real use. Zendesk defines a minimum viable product as the most basic feature set you can offer while still attracting new customers. Keep the scope that small, so a failed test costs little.
The strongest signal in the Goarin model is behavioural: 3-5 similar design partners actively using your product. Similarity matters because one unusual customer proves little about a market. Use is also more convincing than praise, because praise costs the customer nothing.
Productboard’s question for this stage is “Have I built something people want?” Once prospects start choosing or rejecting you, win loss analysis helps explain the reasons behind each decision, which tells you whether to adjust the product or the target segment.
Metrics that signal fit
Fit shows up in a mix of qualitative and quantitative signals. ProductPlan’s glossary, citing Andrew Chen, lists NPS, churn rate, growth rate and market share among the quantitative measures, with word of mouth on the qualitative side. It also states plainly that no single set of metrics can tell any business when it achieves product-market fit.
Read the metrics as a group. A good NPS with high churn suggests people like the idea but the product is not holding them. Low churn with slow growth points to a demand problem, not a product one. Matching each metric to the First Round dimensions makes this easier: NPS speaks to satisfaction, growth rate to demand, and payback to efficiency.
For the wider set of subscription KPIs behind these signals, see Top SaaS Metrics in 2026: 7 KPIs Every Founder Needs To Track – Startalliance.
Stage 3: repeatable growth
Stage 3 is where learning turns into a system. First Round calls repeatability the holy grail of the hunt for fit. In the Goarin model, the repeatability stage looks like growing to 5-25 customers with consistent success patterns and Net Revenue Retention approaching or exceeding 100%. The efficiency stage adds sustainable unit economics, with CAC payback under 12 months.
These markers suit software businesses with recurring revenue. If yours is a different model, keep the logic but choose equivalent measures for retention and payback. The common rule is to scale spend only after customers stay and acquisition pays for itself. Adding budget to a leaky product only makes the leak larger.
Deciding whether you have fit
You have fit when demand, satisfaction and efficiency all hold together and the pattern repeats with new customers. Because fit moves in levels, the practical question is which level you have reached and what would raise it.
Check three things together. First, whether similar customers use the product without hand-holding. Second, whether retention and payback hold as the customer count grows. Third, whether demand is widening beyond your first contacts, which is the direction First Round’s extreme level describes. If one of these fails, return to the stage that owns it instead of pushing on to the next.
Frequently asked questions
What is an MVP in the context of product-market fit?
An MVP is the most basic feature set you can offer while still attracting new customers, according to Zendesk. It exists to test demand cheaply during Stage 2.
Can I run discovery with people who do not use my product yet?
Yes. Zendesk suggests surveying your competitors’ customers to learn what they like, what they dislike and what would make them leave.
What can I change if fit looks weak?
First Round lists four levers for changing who your marginal customer is: your persona, your problem, your promise or your product. Change one at a time so you can see which one moved the results.
Is word of mouth a reliable sign of fit?
It is a strong qualitative signal. Sam Altman’s test, as summarised by Productboard, is that users love the product so much they spontaneously tell other people to use it. Pair it with retention and payback data rather than relying on it alone.
